pdn: dont rely on just the intersection for via insertion, check actual layer shape if initial ensure does not yield valid vias

[gadfort]

Summary

If the overlap / intersection does not allow for any enclosure, PDN would just reject that via, this change ensures we add back the enclosure from the actual shape if possible when the via enclosure is violated.

Type of Change

• Bug fix

Impact

Allows for vias to be inserted with thinner metals

Verification

• I have verified that the local build succeeds (./etc/Build.sh).
• I have run the relevant tests and they pass.
• My code follows the repository's formatting guidelines.
• I have included tests to prevent regressions.
• I have signed my commits (DCO).

[gemini-code-assist]

Code Review

This pull request introduces the calculation and application of spare enclosures for bottom and top layers during via generation, updating enclosure checks and disabling caching when spare enclosures are applied. The feedback suggests simplifying the code in ViaGenerator::determineRowsAndColumns by removing redundant duplicate variables (double_enc_x_lower/double_enc_x_upper and double_enc_y_lower/double_enc_y_upper) that have identical values and can be merged back into single variables.

[maliberty]

The whittle test fails

[2026-08-20T17:31:05.227Z] [whittle] Phase: Nets, Step 39, Insts: 6, Nets: 0, .odb: 1MB (was 0MB), Elapsed: 8m
[2026-08-20T17:31:05.227Z] Traceback (most recent call last):
[2026-08-20T17:31:05.227Z]   File "/home/jenkins/agent/workspace/OpenROAD-Public_PR-11193-merge-2-bazel-ci/etc/whittle.py", line 516, in <module>
[2026-08-20T17:31:05.227Z]     main()
[2026-08-20T17:31:05.227Z]   File "/home/jenkins/agent/workspace/OpenROAD-Public_PR-11193-merge-2-bazel-ci/etc/whittle.py", line 512, in main
[2026-08-20T17:31:05.227Z]     whittler.debug()
[2026-08-20T17:31:05.227Z]   File "/home/jenkins/agent/workspace/OpenROAD-Public_PR-11193-merge-2-bazel-ci/etc/whittle.py", line 470, in debug
[2026-08-20T17:31:05.227Z]     current_err, cuts = self.perform_step(cut_index=j)
[2026-08-20T17:31:05.227Z]                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[2026-08-20T17:31:05.227Z]   File "/home/jenkins/agent/workspace/OpenROAD-Public_PR-11193-merge-2-bazel-ci/etc/whittle.py", line 384, in perform_step
[2026-08-20T17:31:05.227Z]     start = num_elms * cut_index // cuts
[2026-08-20T17:31:05.227Z]             ~~~~~~~~~~~~~~~~~~~~~^^~~~~~
[2026-08-20T17:31:05.227Z] ZeroDivisionError: integer division or modulo by zero

I don't know how this relates to this change.

[gadfort]

@maliberty I saw that too, I can't imagine that it's related, I didn't look that closely at the test. It's possible retriggering it will clear it.

[gadfort]

@maliberty the rerun failed in new ways. Seems like the CI is having some issues.